CtrlK
BlogDocsLog inGet started
Tessl Logo
PluginPublic

ainativedev/aidevcon-2026-nyc

AI Native DevCon 2026 New York — practical agent skills grounded in conference talks

Quality

86%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Overview
Quality
Evals
Security
Files

application-template.mdskills/talk-finster-continuous-delivery/references/

Delivery constraint and thin-slice worksheet

This is an implementation worksheet adapted from Finster's talk. Complete it using one actual change or a clearly labeled hypothetical example.

1. Trace the change

  • Intended user or business outcome:
  • Request received / implementation started / review ready / validation complete / production available:
  • Source for each timestamp:
  • Waiting, rework, and active work:
  • Largest accumulated batch and why it grew:
  • Current release or operational constraints:

State the most likely constraint and the evidence that supports it. If several stages have missing measurements, keep the conclusion provisional. Identify the smallest observation that would distinguish competing explanations.

2. Specify one slice

  • User-visible behavior delivered:
  • What is explicitly excluded:
  • Acceptance example with input, action, and observable result:
  • Boundary or failure example:
  • Dependency that could prevent independent delivery:
  • How product and engineering understanding are reconciled:

Check that the slice is small enough to validate and deliver without collecting unrelated work. Explain which later slices remain useful and whether they truly depend on this one.

3. Design checks and ownership

ConcernDeterministic check or focused judgmentRequired evidenceResponse to failureOwner
Acceptance behavior
Regression risk
Operational fitness

Name a known bad example each important check should reject. Separate a proposed test from a result already observed. If working with legacy code, record current behavior, domain validation, mutation evidence, and only then the proposed refactoring boundary.

4. Run the learning loop

Proposed improvement: Baseline and observation window: Expected reduction in waiting or rework: Quality measure that must not deteriorate: What would falsify the hypothesis: Escaped-defect feedback into the pipeline: Next smallest action within current authorization:

Keep code-generation speed separate from end-to-end delivery improvement.

skills

talk-finster-continuous-delivery

tile.json